Per-MW pricing, regional variance, and cost drivers for owners scoping hyperscale & AI builds.
Salary benchmarks across the 14 mission-critical disciplines.
I wouldn’t launch a data center PM search until the role, fees, screening rules, and owners are approved in writing. Start by defining the job as physical construction delivery - not IT implementation and requesting 2–3 comparable search records plus references.
Here’s what I’d check before giving an agency the go-ahead:
My final rule: <u>no launch with unresolved material role requirements or contract terms</u>. The Project Executive gives final approval.
Approve the recruiting agreement before sourcing begins. The signed contract should name the engagement model: contingent direct hire, retained or engaged search, or RPO vs. in-house recruitment for project-based needs. Don’t rely on the pricing label to explain what you’re buying. Spell out the agency’s deliverables, such as sourcing, technical screening, interview coordination, and offer support. For retained work, tie installments to defined milestones. For RPO or project-based work, define recruiting capacity, the service period, reporting, and any per-hire charges.[3][4][10] Settling these details upfront helps prevent fee disputes once sourcing starts.
For a contingent fee, confirm exactly what triggers payment: offer acceptance, a signed employment agreement, or the candidate’s start date. Ask for a one-page fee schedule covering the fee, included compensation, minimum fee, invoice date, payment deadline, taxes, and pass-through costs. List base salary separately from bonuses, equity, relocation, allowances, and other compensation. State how later compensation changes affect the invoice, and require written approval for candidate travel and other pass-through costs.[5][6][9]
Define a candidate introduction by the candidate’s name, submission date, role, and consent. Set a clear ownership period, and state whether a later hire by another department or affiliate triggers a fee. Track submissions through one contact or an applicant-tracking system to monitor fees and avoid duplicates. Require notices of duplicate submissions or prior contact within two business days, supported by timestamps and existing records.[3][7][8][12] Limit access to candidate data and document the duties for its transfer, use, retention, and deletion.
Spell out termination notice, cancellation charges, and obligations for candidates introduced before the agreement ends. State what remains payable if the search pauses or the requisition changes. Require written approval for changes to scope, compensation, location, or travel requirements. Also clarify whether unused fees can transfer to another approved role.
Replacement terms need the same level of detail as payment terms. If the agreement offers a 90-day replacement credit, confirm when coverage starts, which departures qualify, and whether pre-start withdrawals or failed checks are covered. Specify exclusions, the notice deadline, required invoice status, credit value, expiration, and whether the credit applies only to the same role or another agreed requisition. A replacement credit is conditional coverage - not a refund or guarantee.[11][12]
Data Center PM Candidate Scorecard Weights
Once the agreement is signed, put the search plan in writing. Use it to guide market coverage, screening, and reporting.
Approve the search plan before accepting candidate submissions. Ask the agency to specify where it will source candidates and when it will deliver them. Map Northern Virginia, Phoenix, and Dallas–Fort Worth, along with backup and relocation markets.
Define target employers, employer types, project phases, adjacent titles, disciplines, and seniority. A “data center” label alone isn’t enough. Record source dates, sample size, and data limits, and keep market data separate from self-reported pay.
Assign an owner and a required output to each checkpoint: intake approval, market map, sourcing launch, calibration review, finalist review, reference checks, and offer acceptance.
Agree on when to revisit pay or requirements. Triggers might include repeated compensation objections, limited mobility, or clusters of matches from adjacent job titles. Require outreach and response counts to back up recommendations - not job-board volume alone.
The market map should keep candidate evaluations tied to the same market and scope assumptions.
Separate pass/fail minimums from scored skills. Use the approved brief to screen for construction delivery experience. If data center experience isn’t mandatory, include comparable technical facilities. Also check site access, travel, work authorization, licenses, certifications, security eligibility, and education. Apply the same job-related criteria to every candidate.
Build the scorecard around construction delivery, MEP coordination, commissioning, schedule recovery, and safety. As a starting point, weight lifecycle delivery at 20%; schedule recovery, budget control, and MEP coordination at 15% each; commissioning talent, procurement and subcontractor management, and safety at 10% each; and stakeholder communication at 5%. Adjust the weights to fit the assignment before screening begins.
Rating scale: 1 = no relevant evidence; 2 = limited exposure; 3 = independently managed comparable scope; 4 = led complex scope with measurable results; 5 = repeatedly led comparable mission-critical delivery and resolved major risks.
OPM identifies job-related structured interviews built around competencies from job analysis as highly valid and reliable.[15][16]
Apply these standards to every candidate record before any interview.
Require a completed profile - not just a résumé - so candidates can be compared consistently. Use the table below as a completeness checklist.
Include facility type, project scale, phase, responsibilities, procurement scope, systems exposure, and energization history. Record relevant tools, such as Primavera P6 and Procore, and explain how the candidate used them.
Label evidence as candidate-reported, recruiter-observed, employer-verified, or reference-verified. Being associated with a project doesn’t prove ownership.
Use these records for reporting, scheduling, and feedback.
Use the approved scorecard and candidate profile to keep reporting, scheduling, and offers aligned.
Once screening is standardized, use the same funnel definitions across every search.
Set written SLAs at launch: weekly for retained or urgent searches and biweekly for standard searches. Name an owner for every handoff.[18][19][20][24]
Reports should include outreach totals, screens, submissions, interview status, rejection reasons, market feedback, blockers, and next steps. Give each item one owner and a due date.[21][23]
Define stages before you count them: contacted means outreach sent; engaged means a response to discuss the role; qualified means screening minimums met; submitted means a complete profile sent; and interviewed means an interview completed. Report weekly and cumulative counts for contacted, engaged, qualified, submitted, interviewed, offered, accepted, and started. Keep timestamps and stage definitions consistent across searches.[20][21]
With funnel reporting in place, name the people responsible for moving candidates through interviews and offers.
Route all feedback through the hiring-team owner named in the search brief. A starting SLA allows 72 hours for submission review, 72 hours to confirm an approved interview, and 48 hours to submit post-interview feedback. State whether these are elapsed or business hours, when each clock starts and stops, and how candidate availability or access delays affect timing. Count scheduling only when the interview is confirmed. Escalate missed deadlines to the feedback owner and agency lead. Send repeated delays to the executive sponsor.[19][24]
Name owners for calendar holds, candidate prep, travel, reminders, and status updates. The hiring team owns panel availability and scorecards; site or security staff own access approvals. Only the authorized compensation owner approves pay, relocation, and offer exceptions. Explicitly assign offer issuance, negotiation support, acceptance tracking, and start-date confirmation. For RPO, document these duties in the statement of work.[21]
Keep a dated change log for changes to compensation, location, scope, travel, reporting line, and start date. Include the old and new terms, approver, effective date, and impact on fees or milestones. Before closing or restarting a search, reconcile candidate ownership, duplicate submissions, confidentiality, record-retention rules, outstanding invoices, and candidate notice duties.[13][14][17][22]
For a replacement request, record the start date, departure date, notice date, named owners, coverage status, and agreed restart date. Get written approval for any change that affects replacement coverage.[13][14][17][22]
Use these final checks to decide whether the partner can launch. Assign one of three statuses: approved, open - owner assigned, due date set, or blocked - do not launch.
Then check the reporting and replacement controls. Reporting and interview service levels must have named owners. Replacement terms must spell out the coverage period, qualifying departures, exclusions, notice deadline, and remedy.
Log every open item in a shared launch tracker, including its business impact, required action, owner, and due date. The Project Executive is the final approver. Do not launch until all material in-demand data center roles and requirements and contract terms are closed.
Choose a pricing model based on hiring volume, role complexity, and the cost of project delays. Success-based pricing works well for a single, mission-critical role and requires no upfront fees. For multiple hires, a monthly retainer paired with lower success fees can make costs more predictable.
Compare total cost per successful hire, not just fee percentages. Factor in how long the role stays vacant, internal interview costs, and project schedule delays. Make sure your agreement includes a 90-day replacement guarantee.
Verify day-to-day responsibilities - not just job titles. Check what the candidate personally owned: cost and schedule management, change and turnover deliverables, commissioning and testing leadership, or MEP integration.
Ask the recruiter for a 12–24-month placement log, anonymized sample intake forms, and an anonymized candidate scorecard. The scorecard should separate must-have ownership from preferences that can be taught and identify disqualifiers. Confirm that the recruiter can explain each candidate’s personal responsibilities and provide retention data at 6, 12, and 24 months to validate screening quality [1][2].
Change course on a stalled search as soon as the current approach stops producing results. Step in immediately if your recruiting partner can’t provide weekly reports on pipeline stages, blockers, and search risks [1].
Act on market feedback early when your pay bands are too narrow or your target geography has too few candidates [1]. If you wait until the unfilled role holds up a project, you’re already behind before interviews begin [2].
